Title :
QRP03-1: An Epidemiological Model for File-Sharing with BitTorrent-like Incentives: The Case of a Fixed Peer Population
Author :
Kesidis, G. ; Jin, Y. ; Mortazavi, B. ; Konstantopoulos, T.
Author_Institution :
Depts of CSE & EE, Pennsylvania State Univ., University Park, PA
fDate :
Nov. 27 2006-Dec. 1 2006
Abstract :
Bit Torrent is a popular peer-to-peer file-sharing network that employs transaction-level incentives, i.e., typically two files are swapped for each transaction between peers. Also, larger files are segmented into "chunks" that are the subject of individual transactions. In this paper, we give a simple deterministic stratified epidemiological models of the dissemination of a single popular file in peer-to-peer file-sharing networks that employ BitTorrent-like incentives. We then use this simple model to evaluate the effect of these incentives by comparison with a system does not segment files and always involves only a single file transfer per transaction, i.e., involves a client-peer and server- peer.
Keywords :
peer-to-peer computing; BitTorrent-like incentives; client-peer; deterministic stratified model; fixed peer population; peer-to-peer file-sharing network; server-peer; single file transfer per transaction; transaction-level incentives; Costs; Intelligent networks; Internet; Markov processes; Network servers; Peer to peer computing; Robustness; Telecommunication traffic; Traffic control; Web server;
Conference_Titel :
Global Telecommunications Conference, 2006. GLOBECOM '06. IEEE
Conference_Location :
San Francisco, CA
Print_ISBN :
1-4244-0356-1
Electronic_ISBN :
1930-529X
DOI :
10.1109/GLOCOM.2006.430